There is much frenzied debate about our impending AI doom but I take the Dad's Army and Corporal Jones "don't panic" approach to these pronouncements and focus for now on the many positives and some of my recent discoveries of features you might find useful …
1.PROJECTS — Give your AI a memory for the job
If you repeatedly use AI for the same area of work, stop starting from a blank chat. A Project lets you keep related chats, files and instructions together so the AI has the context it needs from the outset. In ChatGPT, Projects are designed for ongoing work such as writing, research and planning; Claude offers a similar workspace with its own knowledge base and project instructions. Why it helps: less repetition, more context, better continuity.
2.CONNECTIONS — Let AI go to the information
Instead of constantly downloading files, copying text and re-uploading it, connect ChatGPT to services you already use. Depending on the app and your permissions, it can search connected information, bring relevant context into a conversation and sometimes complete approved actions. Why it helps: less admin, fewer stale copies, and AI working closer to where your real information already lives.
3.DEEP RESEARCH — Make it investigate, not just answer
For anything important, switch from normal chat to Deep Research in ChatGPT or Research in Claude. These modes conduct multiple searches, follow questions as they emerge, compare evidence and return an answer with sources you can check. Why it helps: better for decisions, comparisons and evidence-based work than a quick one-shot answer.